Perspectives

Hypercalcemia of Malignancy RevisitedGregory R. MundyDivision of Endocrinology and Metabolism, University of Texas Health Science Center, San Antonio, Texas 78284

In the three years since the last Perspectives article in the Jour-nal of Clinical Investigation on this topic (1), there have beenmajor advances in our understanding of the heterogeneousmechanisms responsible for the hypercalcemia of malignancy.Reports of many of these advances have appeared in the pagesof the Journal of Clinical Investigation. These advances in-clude discovery and characterization of the PTH-related pro-tein (PTH-rP)' produced by many solid tumors, demonstra-tion of tumor-derived transforming growth factor-alpha(TGF-alpha) as a bone-resorbing factor that can cause hyper-calcemia, identification of lymphotoxin as thie major mediatorof bone destruction in myeloma, and definitive demonstrationof increased renal tubular reabsorption of calcium in hyper-calcemia associated with solid tumors.

One concept that has not changed in these last three yearsis that the mechanisms responsible for hypercalcemia of malig-nancy are heterogeneous. Different patterns of disturbances incalcium homeostasis occur in different types of malignancy. Asimple diagram illustrating some of these different patterns isshown in Fig. 1. In all of the examples provided, increasedbone resorption is an important feature. However, the impor-tance of the net fluxes of calcium that occur across the othertwo major organs that guard calcium homeostasis (gut andkidney) depend on the type of tumor.

PTH-related protein (PTH-rP) and solid tumorsIn the last 12 months, a succession of papers has described aprotein isolated from solid tumors that has significant se-quence homology in the NH2-terminal region to PTH. It wasfirst suggested by Albright almost 50 years ago that hypercal-cemia of malignancy could be related to PTH or a PTH-likepeptide, but it was not appreciated just how PTH-like thisfactor was until it was purified from a variety of tumor celllines and molecularly cloned (2-6). This protein is composedof 141 amino acids, with 8 of the first 13 residues identical tothose of human PTH. Since the NH2-terminal region of thePTHmolecule is responsible for biologic activity and bindingto the PTH receptor, it is not surprising that PTH-rP mimicsthe effects of PTH by binding to and activating the PTH re-

Address reprint requests to Dr. Gregory R. Mundy, Division of Endo-crinology and Metabolism, University of Texas Health Science Center,7703 Floyd Curl Drive, San Antonio, TX 78284.

Receivedfor publication 18 February 1988 and in revisedform 5April 1988.

1. Abbreviations used in this paper: PTH-rP, PTH-related protein;TGF, transforming growth factor; TNF, tumor necrosis factor.

ceptor (3). Synthetic peptides that correspond to the NH2-ter-minal end of PTH-rP have now been tested in vitro and invivo; these peptides appear to have effects essentially identicalto those of PTH on the kidney and bone (7-12). Syntheticpeptide analogues of PTH-rP enhance osteoclastic bone re-sorption in vitro (7, 9, 10), enhance renal tubular calciumreabsorption (7, 9), and cause hypercalcemia in vivo (8-10).Differences in potency with PTH have been suggested (7),although these have not been found by other workers (8-10).

These observations provoke obvious questions about thenormal physiologic role of this protein and its relationship toPTH and to the PTH receptor. The immediate question iswhether this protein explains all of the manifestations of hy-percalcemia of malignancy in patients with solid tumors.Based on current data, this is unlikely. The syndrome of hu-moral hypercalcemia of malignancy associated with solidtumors is quite distinct from that of primary hyperparathy-roidism, and cannot be explained simply by what is known ofthe effects of synthetic peptides of PTH-rP on kidney, bone,and gut. For example, osteoclastic bone resorption is muchmore prominent in humoral hypercalcemia of malignancythan in primary hyperparathyroidism, bisphosphonate ther-apy is more effective in lowering the serum calcium, and avail-able information suggests rates of bone formation are reduced.There are also differences in the renal tubular function of thetwo syndromes. Patients with primary hyperparathyroidismhave normal or increased 1,25 dihydroxyvitamin D produc-tion, which is probably a consequence of increased PTHandlowered ambient phosphate concentrations. However, in thehypercalcemia of malignancy, even when PTH-rP is producedby the tumor, 1,25 dihydroxyvitamin D production is usuallysuppressed (13) and patients tend to be alkalotic, rather thanhyperchloremic and mildly acidotic, as is the case in primaryhyperparathyroidism (14). How can these differences be ex-plained? One possibility is that the full-length molecule be-haves differently from the synthetic peptide. However, it isunlikely that opposite effects (e.g., decreased bone formationand decreased 1,25 dihydroxyvitamin D production) will beexplained in this way. A more likely possibility is that theeffects of the PTH-rP on target organs are modified by otherfactors, such as TGF-alpha, IL- 1, and tumor necrosis factor(TNF). These factors are all more powerful bone-resorbingfactors than PTH in vivo (9, 15) and in vitro (16). Each in-creases plasma calcium (15, 17, 18). Evidence for the role ofthese factors was discussed in more detail in the earlier Per-spectives article on this topic (1). Since these factors are oftenproduced by tumors, which also produce PTH-rP, hypercalce-mia in these circumstances is probably due to the combinedeffects of PTH-rP and these factors on target organs. Thesefactors act synergistically with PTH (and probably also withPTH-rP) to stimulate bone resorption in vitro (19) becausetheir major effect is to stimulate proliferation of osteoclast

Hypercalcemia of Malignancy Revisited I

J. Clin. Invest.© The American Society for Clinical Investigation, Inc.0021-9738/88/07/0001/06 $2.00Volume 82, July 1988, 1-6

Myeloma Lung Cancer

Ca++ A~ Ca++ A

CAMPN

Lymphoma Breast Cancer

=CaIL w++A

Ca++ -hCa~

Figure 1. The heterogeneous pattern of abnormalities in calcium ho-meostasis found in patients with the hypercalcemia of malignancy.In myeloma, increased bone resorption is associated with decreasedglomerular filtration. In patients with solid tumors such as lungcancer, increased bone resorption and increased renal tubular cal-cium reabsorption occur, and are associated with increased nephro-genous cAMP. In patients with breast cancer, increased bone resorp-

tion and increased renal tubular calcium reabsorption also occur, butthis tumor is not usually associated with increased nephrogenouscAMP, and is almost always associated with metastatic bone disease.In lymphomas, increased bone resorption occurs together with, insome cases, increased gut absorption of calcium. The potentialmechanisms responsible for these derangements are described in thetext. Reproduced from Kidney International, 1987. 31:142-155, pub-lished with permission.

progenitors (20), whereas PTHaffects cells that are later in theosteoclast lineage. However, the effects of PTH (and presum-

ably PTH-rP) on some target cells may also be substantiallyimpaired by these factors. For example, TGF-alpha, IL- 1, andTNF all decrease adenylate cyclase responsivity to PTH incultured osteoblastic cells (21), and it is likely they will havethe same effects on these responses to PTH-rP.

Another question is whether the complete protein will haveother functions, possibly by interacting with another receptordistinct from that of PTH. The answer to that question may

have to await availability of preparations of the recombinantprotein. The PTH-rP shares 8 of the first 13 amino acids withPTH, but thereafter there is no sequence homology. AlthoughPTH (1-84) behaves similarly to PTH (1-34) in biologicalassays, PTH-rP has a basic amino acid carboxyl terminus of 57residues that could well mediate other biological effects.

According to the data of Moseley and co-workers (22),PTH-rP is produced very commonly by many solid tumors.For example, the majority of squamous cell cancers of the skinand normal skin cells contain PTH-related peptide as detected

by immunofluorescent techniques. Since PTH-rP is not asso-ciated with hypercalcemia in these cases, it is likely that thetumors do not secrete sufficient amounts of the PTH-rP todisrupt calcium homeostasis. Could PTH-rP have a normalphysiological function on calcium or phosphate homeostasis?This does not appear likely in postnatal life, unless the normaltissue source is the parathyroid glands. Patients with parathy-roid gland disease or ablation have hypocalcemia and hyper-phosphatemia, indicating that a normal tissue source outsidethe parathyroids does not produce enough PTH-rP to preventthe features of hypoparathyroidism. It is conceivable, however,that PTH-rP is an alternative form of PTH produced in theparathyroid glands themselves. If so, it could be responsible forsome cases of primary hyperparathyroidism. There are a fewpatients with primary hyperparathyroidism in whom in-creased serum-immunoreactive PTH concentrations cannotbe detected. Another possibility is that PTH-rP, like someother tumor products, represents a fetal hormone, affectingcalcium homeostasis before birth, and preliminary evidencesuggests that PTH-rP derived from the fetal parathyroid glandsmay influence placental calcium transport in fetal lambs (23).

One point to emphasize is that although the PTH-rP isfound frequently in solid tumors, there is no evidence as yetthat PTHitself is produced by any tumors except those arisingfrom the parathyroid glands (24). It has been suggested thatassays used in earlier studies showing increased serum-immu-noreactive PTH in patients with malignancy (25) were in factrecognizing PTH-rP. This does not explain how these resultswere also obtained in patients in whom PTH-rP is not pro-duced, such as patients with myeloma. Moreover, the similar-ities in the molecules are not in the regions best recognized bythese older assays. An alternative explanation for the spuriousimmunoassay results may be that the suppressed parathyroidglands release immunoreactive but biologically inert PTHfragments.

The relationship of the PTH-rP gene to that for PTH itselfshould soon be known. It has recently been reported by Man-gin et al. (6) that PTH-rP is encoded by human chromosome12, whereas the PTH gene has previously been mapped tochromosome 1 1. PTH-rP thus cannot be the result of alternatesplicing of the PTHgene. Mangin et al. (6) do however reportmultiple forms (at least three) of PTH-rP mRNA.One of theircDNAs may encode an alternately spliced message that couldgive rise to a second PTH-rP, with a novel carboxy-terminalportion, thus bearing the same relation to PTH-rP that calci-tonin gene-related protein does to calcitonin (perhaps a PTH-rP gene-related protein). It is likely that both PTHand PTH-rParose from a single ancestral gene that duplicated and sepa-rated during evolution, such as is the case with epidermalgrowth factor and TGF-alpha. These answers should be clearerwhen the complete gene sequence is published.

Lymphotoxin and myelomaNew proteins are being identified not only in solid tumorsassociated with hypercalcemia, but also in the hematologicmalignancies. In myeloma, it has long been known that themalignant plasma cells produce a local cytokine that stimu-lates osteoclast activity (26). Recently, the immune cell prod-ucts lymphotoxin, TNF, and IL- 1 have each been identified aspotent stimulators of osteoclastic bone resorption in vitro (27,28). Similar activity is produced by normal activated immunecells (for review and bibliography, see reference 1). Although

2 G. R. Mundy

AM

the majority of bone-resorbing activity produced by normalactivated leukocytes is IL-1 beta (29), it is now apparent thatthe major bone-resorbing cytokine expressed by those humanmyeloma cells so far studied is lymphotoxin (also called TNFbeta). The bone-resorbing activity produced by five culturedmyeloma cell lines in vitro was substantially inhibited by neu-tralizing antibodies to lymphotoxin (30). Moreover, lympho-toxin causes hypercalcemia when infused or injected in vivo(30). Lymphotoxin is normally produced by activated T lym-phocytes. It binds to the same receptor as TNF, with which itshares identical biological properties. Its biological propertiesalso overlap with many of those of IL- 1. This group of mole-cules may be responsible for other ectopic hormone syn-dromes in patients with cancer, including cachexia, hypertri-glyceridemia, and probably anemia due to suppression oferythropoiesis. These cytokines not only stimulate prolifera-tion of osteoclast progenitors (31), but also activate the maturemultinucleate cells, albeit indirectly (32). However, like PTH-rP produced by solid tumors, production of lymphotoxin bymyeloma cells does not alone account for the hypercalcemicsyndrome in myeloma. Not all of the bone-resorbing activityproduced by myeloma cells can be blocked by antisera to lym-photoxin (30), and it appears likely that other bone-resorbingfactors that remain to be identified are also produced by my-eloma cells.

Interestingly, human myeloma cells constitutively expressmRNAfor both lymphotoxin and TNF(30). However, TNF isnot released by the cells. Presumably, there is inhibition oftranslation or secretion of the protein.

Production of potent bone-resorbing cytokines by neoplas-tic immune cells may represent an aberration of the mecha-nisms that regulate normal osteoclastic bone resorption. Boneis normally remodeled in discrete units or packets throughoutthe skeleton, and the cellular events involved in remodelingare likely to be controlled by local factors produced in themicroenvironment of remodeling surfaces. Normal immunecells that produce potent osteotropic cytokines such as lym-photoxin, IL- 1, TNF, gamma-IFN, and TGFbeta may occupya pivotal role in the regulation of normal bone remodeling.

Relative roles of the kidney and bone in humoralhypercalcemiaThe major organs involved in disruptions of calcium homeo-stasis in- hypercalcemia of malignancy are bone and kidney.Their relative importance is controversial. Bone resorption hasalways been considered the major cause by most Americaninvestigators, but that position has been strongly challengedrecently. European investigators have long considered that themajor organ responsible for hypercalcemia in primary hyper-parathyroidism is the kidney, which they believe to play animportant and underappreciated role in the hypercalcemia ofmalignancy as well. The earliest observations were based onclinical studies (33), which showed that renal calcium excre-tion was usually less than expected for the degree of hypercal-cemia. Most patients with hypercalcemia of malignancy (my-eloma is a notable exception) have increased renal tubularreabsorption of calcium, similar to those with primary hyper-parathyroidism. These observations have been criticized onthe grounds that impaired volume status could promote cal-cium reabsorption, and this could be the explanation for im-paired calcium excretion. However, Bonjour and his co-workers recently showed in the rat Leydig cell tumor model of

the humoral hypercalcemia of malignancy, that reabsorptionof calcium was increased in the renal tubules in circumstancesin which sodium and volume status were carefully controlled.They then showed that the radioprotective agent WR-2721,which inhibits renal tubular calcium reabsorption, loweredserum calcium at the same time as it increased renal calciumexcretion (34). The only conclusion from these observationscan be that renal tubular calcium reabsorption plays an im-portant role in the pathophysiology of the humoral hypercal-cemia in this model. This conclusion has been strengthenedrecently by observations that synthetic peptides of PTH-rP(1-34) increase renal tubular calcium reabsorption in vivo (9),and Leydig tumor cells are known to produce a PTH-rP (35).A different mechanism for increasing renal tubular calciumreabsorption has been proposed by Bijvoet and his colleagues(36). This group has gained enormous experience over theyears with the agent 3 amino-l-hydroxypropylidine-1,1-bis-phosphonate (APD), a bisphosphonate that specifically in-hibits bone resorption. They argue that since APDis remark-ably effective in lowering serum calcium in hypercalcemia ofmalignancy (> 95% of patients respond), increased bone re-sorption must be the major pathogenetic mechanism responsi-ble for hypercalcemia. Bijvoet and co-workers also argue that acomponent of increased renal tubular calcium reabsorptionexists, but suggest that this occurs not as a consequence of atumor factor but rather because of the natriuretic effect ofhypercalcemia that leads to volume depletion. As a conse-quence, proximal tubular reabsorption of calcium, sodium,and fluid are enhanced. Sodium loading restores extracellularfluid volume and reduces enhanced reabsorption of sodiumand calcium in the proximal tubules, which is due to thismechanism. However, when urinary sodium excretion exceeds4 mmol/liter glomerular filtrate (fractional excretion of so-dium 3%), the effects of a humoral factor on calcium reab-sorption in the distal tubules will also be obscured (37), be-cause excessive natriuresis markedly enhances distal deliveryof calcium so that reabsorptive capacity at that site is exceeded(37). Under conditions of vigorous sodium loading, the effectsof a humoral factor thus may not be apparent and increasedrenal tubular calcium reabsorption will be lowered whether itis due to a humoral factor or to volume depletion caused bythe natriuretic effect of hypercalcemia.

The kidney is also important in the hypercalcemia of my-eloma. In patients with myeloma who develop hypercalcemia,there is almost always some fixed impairment of renal func-tion (38) that may have multiple causes. In this case, hypercal-cemia results from increased bone resorption associated withdecreased glomerular filtration and decreased capacity of thekidney to clear the calcium load from the extracellular fluid.

Increased osteoclastic bone resorption is obviously an im-portant mechanism in the pathophysiology of hypercalcemiaof malignancy, since hypercalcemia is improved by drugs thatspecifically inhibit bone resorption and increased osteoclastactivity is usually a prominent morphologic abnormality.However, Ralston and co-workers (39) have recently ques-tioned whether this increased bone resorption is important inearly hypercalcemia and whether it is factor mediated. Usinghistomorphometric studies, they argue that hypercalcemia can

occur in patients with malignancy without detectable changesin bone morphology and specifically osteoclastic resorption.Their suggestion is that the tumor factors act on the kidney topromote renal tubular calcium reabsorption, and bone resorp-

Hypercalcemia of Malignancy Revisited 3

tion becomes important only later in the disease when thepatient is sicker and immobile. It should be appreciated thatthe morphologic changes in bone in patients with malignancyand hypercalcemia are not those seen in primary hyperpara-thyroidism, but are consistent with those seen in early stages ofimmobilization. For example, osteoclastic bone resorption ismore prominent, subperiosteal resorption is extremely rare,and there is no evidence of increased bone formation or boneturnover, which is characteristic of florid cases of primary hy-perparathyroidism. Although immobilization may play a rolein promoting osteoclastic bone resorption, it is not likely to bethe only or even major mechanism for bone resorption. Fullymobile animals carrying hypercalcemic tumors have promi-nent osteoclastic bone resorption. Moreover, in vitro studiesusually show that tumors associated with hyperealcemia se-crete factors that are potent stimulators of osteoclastic boneresorption.

Both increased bone resorption and impaired renal cal-cium excretion are clearly important in the pathophysiology ofhypercalcemia, and their combined effects on serum calciumare probably synergistic. For example, a patient with substan-tially increased bone resorption alone may remain normocal-cemic since renal calcium excretion will be enhanced, becausePTH is suppressed. However, if a tumor secretes factors that inaddition to stimulating bone resorption also inhibit the frac-tional excretion of calcium, hypercalcemia will occur morereadily. Questioning whether bone resorption or renal tubularcalcium reabsorption are mediated by factors or other mecha-nisms is useful, but the important issue is the relative impor-tance of bone resorption compared with renal calcium excre-tion, and whether therapy should be tailored to the pathophys-iologic mechanism. Unfortunately, the experimentalapproaches that are currently being used to distinguish therelative importance of bone and kidney in hypercalcemia(such as testing effects of tumor factors on plasma calciumafter total nephrectomy) are too imprecise to provide this in-formation.

Other problems to be solvedBreast cancer and other tumors associated with osteolytic me-tastases. Approximately 25% of unselected hypercalcemic pa-tients with malignant disease have breast cancer (40), makingit the most common hypercalcemic malignancy after lungcancer. Hypercalcemia in breast cancer occurs most com-monly when there is widespread osteolytic bone destructionand in the late phases of the disease. Bone destruction is proba-bly osteoclastic and occurs locally around metastatic tumordeposits, although cultured human breast cancer cells alsohave the capacity to resorb bone (41). The factors responsiblefor this bone resorption are far from clear. The majority ofbreast cancers do not produce PTH-rP (12). In some cases, it ispossible thai prostaglandins are involved as local stimulatorsof osteoclasts (42). Certainly, cultured human breast cancercells release bone-resorbing prostaglandins when exposed toestrogens or antiestrogens (43). However, indomethacin andsimilar prostaglandin synthesis inhibitors have been notor-iously ineffective as therapies for either metastatic bone diseasein breast cancer or hypercalcemia. There are numerous otherosteotropic factors produced by breast cancer cells that couldbe important, including TGFalpha, TGFbeta, and platelet-derived growth factor, each of which can stimulate bone re-sorption. Whether any of these proteins has a role in the bone

destruction or hypercalcemia associated with breast cancer re-mains to be shown. Increased renal tubular calcium reabsorp-tion may also contribute to hypercalcemia in breast cancer(44), although the mechanism responsible is unknown.Clearly, breast cancer is an area ripe for investigation for hy-percalcemia-inducing factors.

Lymphomas. In a few cases of lymphoproliferative disease(Hodgkin's disease, B cell lymphomas, and T cell lymphomas)associated with hypercalcemia, serum 1,25 dihydroxyvitaminD is increased (45). In these patients, gut absorption of cal-cium, when measured, is also increased. The source of 1,25dihydroxyvitamin D is likely to be the tumor cells themselves,since lymphoid cells transformed with the human T cell lym-photrophic virus-type I (which is one cause of adult T celllymphoma) have now been found to metabolize 25 hydroxy-vitamin to 1,25 dihydroxyvitamin D (46). Increased serum1,25 dihydroxyvitamin D is an unusual finding that appears inonly a minority of hypercalcemic patients with lymphoprolif-erative disorders. Even in those in whomincreased 1,25 dihy-droxyvitamin D is present, coincidental mechanisms such asthe production of bone-resorbing cytokines by the lymphoidcells may be at least equally important.

1,25 dihydroxyvitamin D has unique effects on bone re-sorption. It causes differentiation of committed progenitorsinto cells with the characteristics of mature osteoclasts in vitro(47). 1,25 dihydroxyvitamin D has many other immunomo-dulatory effects on T cells and cytokine production in themarrow environment (48). Its production by transformedlymphocytes raises the possibility that it might also be pro-duced by normal lymphocytes in the bone microenvironmentunder appropriate circumstances.

Others. There are other factors and potential mechanismsthat could be responsible for the hypercalcemia of malignancythat have not yet been thoroughly investigated. These includethe production of IL- 1 and other cytokines by solid tumors oreven normal cells in response to the presence of a tumor me-tastasis. As part of the host defense response to the presence ofa tumor, normal immune cells may attempt to limit growth ofthe tumor by releasing cytokines that influence tumor growthand produce other paraneoplastic syndromes such as cachexiaand hyperpyrexia. Possibly hypercalcemia is another, sincethese cytokines can also increase plasma calcium (15, 18, 3p).

A number of tumors have been described that cause leu-kocytosis in association with hypercalcemia. These are fre-quently squamous cell carcinomas of the head and neck, andmany of these tumors produce colony-stimulating factors aldbone-resorbing factors, and cause a hypercalcemia-leukocy-tosis syndrome in nude mice (49). The colony-stimulating fac-tors associated with the leukocytosis syndrome are probablynot identical with the bone-resorbing factors responsible forincreasing osteoclast activity.

Two recent reviews have focused attention on the com-plexities of calcium homeostasis and the oversimplified tradi-tional view of the system as pne of simple reactive regulationby systemic hormones controlled by negative feedback. Parfitt(50) and Staub et al. (51) have drawn attention to the impor-tance of the intrinsic self-oscillatory mechanism that is respon-sible for maintenance of plasma calcium even in the absence ofhormones, and which is mediated by exchange of calciumbetween the bone fluid and the extracellular fluid. This cal-cium exchange depends on the lining cells that cover quiescentbone surfaces, and occurs independent of osteoclastic bone

4 G. R. Mundy

resorption and bone formation. Nothing is yet known of theeffects of tumors or tumor factors on the exchange of calciumbetween the bone fluid and the extracellular fluid.

In conclusion, hypercalcemia of malignancy is an excitingarea that is changing rapidly. Modem techniques in proteinpurification and molecular biology have effectively identifiedtumor-associated proteins that are involved in the syndromesof hypercalcemia associated with malignancy. These studieshave improved our understanding not only of tumor biologyand the production of tumor-associated proteins, but also ourunderstanding of normal calcium homeostasis and the physio-logical factors that regulate osteoclastic bone resorption andrenal tubular calcium handling. Identification of the factorsshould be followed by studies of their relative biological signifi-cance. The most important challenges now facing investigatorsin this area will be to develop adequate biological systems totest the effects of these factors on calcium homeostasis in vivoand to determine the relative importance of their actions onthe various limbs of the homeostatic system.

AcknowledgmentsI am grateful to Nancy Garrett for her help in the preparation of thisreview. I amalso grateful to mycolleagues Lynda Bonewald, BrendanBoyce, John Chirgwin, James Dunn, Gloria Gutierrez, Michael Katz,James Poser, David Roodman, Massimo Sabatini, Peter Smolens,Katherine Tuttle, Ashley Yates, and Toshiyuki Yoneda for helpfuldiscussions.

Some of the data reviewed in this paper have been gathered withsupport from National Institutes of Health grants CA-40035,AR-28149, and RR- 1346.

References21. Mundy, G. R., K. J. Ibbotson, and S. M. D'Souza. 1985. Tumor

products and the hypercalcemia of malignancy. J. Clin. Invest.76:391-395.

2. Suva, L. J., G. A. Winslow, R. E. H. Wettenhall, R. G. Ham-monds, J. M. Moseley, H. Diefenbach-Jagger, C. P. Rodda, B. E.Kemp, H. Rodriguez, E. Y. Chen, P. J. Hudson, T. J. Martin, and W. I.Wood. 1987. A parathyroid hormone-related protein implicated inmalignant hypercalcemia: cloning and expression. Science (Wash.DC). 237:893-896.

3. Strewler, G. J., P. H. Stern, J. W. Jacobs, J. Eveloff, R. F. Klein,S. C. Leung, M. Rosenblatt, and R. A. Nissenson. 1987. Parathyroidhormone-like protein from human renal carcinoma cells. Structuraland functional homology with parathyroid hormone. J. Clin. Invest.80:1803-1807.

4. Moseley, J. M., M. Kubota, H. Diefenbach-Jagger, R. E. H.Wettenhall, B. E. Kemp, L. J. Suva, C. P. Rodda, P. R. Ebeling, P. J.Hudson, J. D. Zajac, and T. J. Martin. 1987. Parathyroid hormone-related protein purified from a human lung cancer cell line. Proc. Nati.Acad. Sci. USA. 84:5048-5052.

5. Stewart, A. F., T. We, D. Goumas, W. J. Burtis, and A. E.Broadus. 1987. N-terminal amino acid sequence of two novel tumor-derived adenylate cyclase-stimulating proteins: identification of para-thyroid hormone-like and parathyroid hormone-unlike domains. Bio-chem. Biophys. Res. Commun. 146:672-678.

6. Mangin, M., A. C. Wegg, B. E. Dreyer, J. T. Posillico, K. Ideda,D. E. Barton, U. Francke, and A. E. Broadus. 1988. Identification of acDNA encoding a parathyroid hormone-like peptide from a human

2. It has not been possible within the context of a Perspectives articleto reference all of the important work in this field, particularly beforethe last several years. A more complete bibliography can be found inreferences 1, 14, and 37.

tumor associated with a humoral hypercalcemia of malignancy. Proc.Natl. Acad. Sci. USA. 85:597-601.

7. Kemp, B. E., J. M. Moseley, C. P. Rodda, P. R. Ebeling, R. E. H.Wettenhall, D. Stapleton, H. Diefenbach-Jagger, F. Ure, V. P. Michel-angeli, H. A. Simmons, L. G. Raisz, and T. J. Martin. 1987. Parathy-roid hormone-related protein of malignancy: active synthetic frag-ments. Science (Wash. DC). 238:1568-1570.

8. Horiuchi, N., M. P. Caulfield, J. E. Fisher, M. E. Goldman, R. L.McKee, J. E. Reagan, J. J. Levy, R. F. Nutt, S. B. Rodan, T. L.Schofield, T. L. Clemens, and M. Rosenblatt. 1987. Similarity of syn-thetic peptide from human tumor to parathyroid hormone in vivo andin vitro. Science (Wash. DC). 238:1566-1567.

9. Yates, A. J. P., G. E. Gutierrez, P. Smolens, P. S. Travis, M. S.Katz, T. B. Aufdemorte, B. F. Boyce, T. K. Hymer, J. W. Poser, andG. R. Mundy. 1988. Effects of a synthetic peptide of a parathyroidhormone-related protein on calcium homeostasis, renal tubular cal-cium reabsorption, and bone metabolism. J. Clin. Invest. 81:932-938.

10. Stewart, A. F., M. Mangin, T. Wu, D. Goumas, K. L. Insogna,W. J. Burtis, and A. E. Broadus. 1988. Synthetic human parathyroidhormone-like protein stimulates bone resorption and causes hypercal-cemia in rats. J. Clin. Invest. 81:596-600.

11. Rodan, S. B., M. Noda, G. Wesolowski, M. Rosenblatt, andG. A. Rodan. 1988. Comparison of postreceptor effects of 1-34 humanhypercalcemia factor and 1-34 human parathyroid hormone in ratosteosarcoma cells. J. Clin. Invest. 81:924-927.

12. Sartori, L., E. C. Weir, A. F. Stewart, A. E. Broadus, M. Man-gin, P. Q. Barrett, and K. L. Insogna. 1988. Synthetic and partially-pu-rified adenylate cyclase-stimulating proteins from tumors associatedwith humoral hypercalcemia of malignancy inhibit phosphate trans-

port in a PTH-responsive renal cell line. J. Clin. Endocrinol. & Metab.66:459-461.

13. Stewart, A. F., R. Horst, L. J. Deftos, E. C. Cadman, R. Lang,and A. E. Broadus. 1980. Biochemical evaluation of patients withcancer-associated hypercalcemia: evidence for humoral and nonhu-moral groups. N. Engl. J. Med. 303:1377-1383.

14. Wills, M. R., and G. K. McGowan. 1964. Plasma-chloridelevels in hyperparathyroidism and other hypercalcaemic states. Br.Med. J. 1:1153-1156.

15. Sabatini, M., B. Boyce, T. Aufdemorte, L. Bonewald, and G.Mundy. 1988. Infusions of recombinant human interleukin-1 alphaand beta cause hypercalcemia in normal mice. Proc. Natl. Acad. Sci.USA. In press.

16. Mundy, G. R. 1987. The hypercalcemia of malignancy. KidneyInt. 31:142-155.

17. Tashjian, A. H., E. F. Voelkel, W. Lloyd, R. Derynck, M. E.Winkler, and L. Levine. 1986. Actions of growth factors on plasmacalcium. J. Clin. Invest. 78:1405-1409.

18. Tashjian, A. H., E. F. Voelkel, M. Lazzaro, D. Goad, T. Bosma,and L. Levine. 1987. Tumor necrosis factor-alpha (cachectin) stimu-lates bone resorption in mouse calvaria via a prostaglandin-mediatedmechanism. Endocrinology. 120:2029-2036.

19. Dewhirst, F. E., J. M. Ago, W. J. Peros, and P. Stashenko. 1987.Synergism between parathyroid hormone and interleukin- 1 in stimu-lating bone resorption in organ culture. J. Bone Min. Res. 2:127-134.

20. Takahashi, N., B. R. MacDonald, J. Hon, M. E. Winkler, R.Derynck, G. R. Mundy, and G. D. Roodman. 1986. Recombinanthuman transforming growth factor alpha stimulates the formation ofosteoclast-like cells in long term human marrow cultures. J. Clin.Invest. 78:894-898.

21. Gutierrez, G. E., G. R. Mundy, R. Derynck, E. L. Hewlett, andM. S. Katz. 1987. Inhibition of parathyroid hormone-responsive ade-nylate cyclase in clonal osteoblast-like cells by transforming growthfactor alpha and epidermal growth factor. J. Biol. Chem. 262:15845-15850.

22. Moseley, J. M., B. E. Kemp, C. P. Rodda, V. P. Michelangeli,P. R. Ebeling, J. A. Danks, L. G. Raisz, and T. J. Martin. 1988.Parathyroid hormone-related protein of malignancy. Biological ac-

Hypercalcemia of Malignancy Revisited 5

tions of synthetic peptides and immunocytochemical localisation.Calcif Tissue Int. 42 (Suppl.): 174.

23. Rodda, C. P., J. A. Heath, P. R. Ebeling, J. M. Moseley, A. D.Care, I. W. Caple, and T. J. Martin. Regulation of fetal calcium metab-olism: evidence for a novel parathyroid hormone-related protein pro-moting placental calcium transport. J. Bone Min. Res. In press.

24. Simpson, E. L., G. R. Mundy, S. M. D'Souza, K. J. Ibbotson,R. Bockman, and J. W. Jacobs. 1983. Absence of parathyroid hor-mone messenger RNAin nonparathyroid tumors associated with hy-percalcemia. N. Engl. J. Med. 309:325-330.

25. Benson, R. C., B. L. Riggs, B. M. Pickard, and C. D. Arnaud.1974. Radioimmunoassay of parathyroid hormone in hypercalcemicpatients with malignant disease. Am. J. Med. 56:821-826.

26. Mundy, G. R., L. G. Raisz, R. A. Cooper, G. P. Schechter, andS. E. Salmon. 1974. Evidence for the secretion of an osteoclast stimu-lating factor in myeloma. N. Engl. J. Med. 291:1041-1046.

27. Bertolini, D. R., G. E. Nedwin, T. S. Bringman, and G. R.Mundy. 1986. Stimulation of bone resorption and inhibition of boneformation in vitro by human tumour necrosis factor. Nature (Lond.).319:516-518.

28. Gowen, M., 0. D. Wood, E. J. Ihrie, M. K. B. McGuire, andR. G. G. Russell. 1983. An interleukin 1 like factor stimulates boneresorption in vitro. Nature (Lond.). 306:378-380.

29. Dewhirst, F. E., P. P. Stashenko, J. E. Mole, and T. Tsuruma-chi. 1985. Purification and partial sequence of human osteoclast-acti-vating factor: identity with interleukin-l beta. J. Immunol. 135:2562-2568.

30. Garrett, I. R., B. G. M. Durie, G. E. Nedwin, A. Gillespie, T.Bringman, M. Sabatini, D. R. Bertolini, and G. R. Mundy. 1987.Production of the bone resorbing cytokine lymphotoxin by culturedhuman myeloma cells. N. Engl. J. Med. 317:526-532.

31. Roodman, G. D., N. Takahashi, A. Bird, and G. R. Mundy.1987. Tumor necrosis factor alpha (TNF) stimulates formation ofosteoclast-like cell (OCL) in long term human marrow cultures bystimulating production of interleukin-l (IL-1). Clin. Res. 35:51 5a.(Abstr.)

32. Thomson, B. M., G. R. Mundy, and T. J. Chambers. 1987.Tumor necrosis factors alpha and beta induce osteoblastic cells tostimulate osteoclastic bone resorption. J. Immunol. 138:775-779.

33. Peacock, M., W. G. Robertson, and B. E. C. Nordin. 1969.Relation between serum and urine calcium with particular reference toparathyroid activity. Lancet. i:384-386.

34. Hirschel-Scholz, S., J. Caverzasio, R. Rizzoli, and J. P. Bon-jour. 1986. Normalization of hypercalcemia associated with a decreasein renal calcium reabsorption in Leydig cell tumor-bearing rats treatedwith WR-2721. J. Clin. Invest. 78:319-322.

35. Rodan, S. B., K. L. Insogna, A. M. C. Vignery, A. F. Stewart,A. E. Broadus, S. M. D'Souza, D. R. Bertolini, G. R. Mundy, andG. A. Rodan. 1983. Factors associated with humoral hypercalcemia ofmalignancy stimulate adenylate cyclase in osteoblastic cells. J. Clin.Invest. 72:1511-1515.

36. Harinck, H. I. J., 0. L. M. Bijvoet, and A. S. T. Plantingh. 1987.Role of bone and kidney in tumor-induced hypercalcemia and its

treatment with bisphosphonate and sodium chloride. Am. J. Med.82:1133-1142.

37. Ralston, S. H., M. D. Gardner, F. J. Dryburgh, R. A. Cowan,and I. T. Boyle. 1986. Influence of urinary sodium excretion on theclinical assessment of renal tubular calcium reabsorption in hypercal-caemic man. J. Clin. Pathol. 39:641-646.

38. Durie, B. G. M., S. E. Salmon, and G. R. Mundy. 1981. Rela-tion of osteoclast activating factor production to the extent of bonedisease in multiple myeloma. Br. J. Haematol. 47:21-30.

39. Ralston, S. H., B. F. Boyce, R. A. Cowan, M. D. Gardner,W. D. Fraser, and I. T. Boyle. 1988. Contrasting mechanisms of hy-percalcaemia in early and advanced humoral hypercalcemia of malig-nancy. CalciJ Tissue Int. 42(Suppl.): 181.

40. Mundy, G. R., and T. J. Martin. 1982. Hypercalcemia of malig-nancy: pathogenesis and treatment. Metab. Clin. Exp. 31:1247-1277.

4 1. Eilon, G., and G. R. Mundy. 1978. Direct resorption of bone byhuman breast cancer cells in vivo. Nature (Lond.). 276:726-728.

42. Galasko, C. S. B., and A. Bennett. 1976. Relationship of bonedestruction in skeletal metastases to osteoclast activation and prosta-glandins. Nature (Lond.). 263:508-510.

43. Valentin-Opran, A., G. Eilon, S. Saez, and G. R. Mundy. 1985.Estrogens and antiestrogens stimulate release of bone resorbing activityby cultured human breast cancer cells. J. Clin. Invest. 75:726-731.

44. Percival, R. L., A. J. P. Yates, R. E. S. Gray, J. Galloway, K.Rogers, F. E. Neal, and J. A. Kanis. 1985. Mechanisms of malignanthypercalcemia in carcinoma of the breast. Br. Med. J. 291:776-779.

45. Breslau, N. A., J. L. McGuire, J. E. Zerwekh, E. P. Frenkel, andC. Y. C. Pak. 1984. Hypercalcemia associated with increased serumcalcitriol levels in -three patients with lymphoma. Ann. Intern. Med.100:1-7.

46. Fetchick, D. A., D. R. Bertolini, P. S. Sarin, S. T. Weintraub,G. R. Mundy, and J. D. Dunn. 1986. Production of 1,25-dihydroxyvi-tamin D by human T-cell lymphotrophic virus-I transformed lym-phocytes. J. Clin. Invest. 78:592-596.

47. Roodman, G. D., K. J. Ibbotson, B. R. MacDonald, T. J.Keuhl, and G. R. Mundy. 1985. 1,25(OH)2 vitamin 13 causes forma-tion of multinucleated cells with osteoclast characteristics in cultures ofprimate narrow. Proc. NatL Acad. Sci. USA. 82:8213-8217.

48. Manolagas, S. C., D. M. Prowedini, and C. Tsoukas. 1985.Interactions of 1,25 dihydroxyvitamin D3 and the immune system.Mol. Cell. Endocrinol. 43:113-122.

49. Sato, K., H. Miumra, D. C. Han, T. Kariuchi, T. Ueyama, H.Ohkawa, T. Okabe, Y. Kondo, N. Ohsawa, T. Tsushima, and K. Shi-zume. 1986. Production of bone resorbing activity and colony stimu-lating activity in vivo and in vitro by a human squamous cell carci-noma associated with hypercalcemia and leukocytosis. J. Clin. Invest.78:145-154.

50. Parfitt, A. M. 1987. Bone and plasma calcium homeostasis.Bone. 8:S1-S8.

51. Staub, J. F., P. Tracqui, P. Brezillon, G. Milhaud, and A. M.Perault-Staub. 1988. Calcium metabolism in the rat: a temporal self-organized model. Am. J. Physiol. 254:R134-R149.

6 G. R. Mundy